Beyond the Peak - herschel.esac.esa.intherschel.esac.esa.int/TheUniverseExploredByHerschel/presentations/10a... · Beyond the Peak Resolved Far-Infrared Spectral Mapping of Nearby

Beyond the PeakResolved Far-Infrared Spectral Mapping

of Nearby Galaxies with SPIRE/FTS

J.D. Smith (Univ. of Toledo)and the BtP team

Thursday, October 17, 13









Beyond the Peak: OverView

~150 hr OT1 program

22 Galaxies, 1 extra-nuclear region, drawn from KingFish survey

High-resolution, fully sampled 200–600µm Mapping spectroscopy (~85% of it)

• Continuum• CO J=4-3 to 13-12• [NII] 205µm• CI 370 & 609 µm• (H20, OH, ...)

Beyond the Peak



Beyond the Peak


JD Smith, Toledo (PI)Lee Armus, SSC/Caltech Pedro Beirao, Paris Obs. Alison Crocker, ToledoKevin Croxall, OSUDanny Dale, WyomingBruce Draine, PrincetonAlberto Bolatto, Maryland Matt Bradford, JPL/Caltech Elias Brinks, Hertfordshire Daniela Calzetti, UMass Brent Groves, MPIALeslie Hunt, INAF, Arcetri Rob Kennicutt, Cambridge Johan Knapen, IAC

Jin Koda, Stony Brook Kathryn Kreckel, MPIAAdam Leroy, NRAO Eric Murphy, SSC/Caltech Eric Pellegrini, ToledoDimitra Rigopoulou, Oxford/RAL Erik Rosolowsky, AlbertaKarin Sandstrom, MPIA Eva Schinnerer, MPIA Paul van der Werf, Leiden Laurent Vigroux, IAP Fabian Walter, MPIA Christine Wilson, McMasterMark Wolfire, Maryland


NGC1097 NGC1266 NGC1377 NGC1482 NGC2798

NGC2976 NGC3077 NGC3351 NGC3521

NGC3627 NGC4254

NGC4321 NGC4536 NGC4569 NGC4631

NGC4736

NGC4826 NGC5055 NGC5457 NGC5713 NGC6946 NGC7331


THANKS SPIRE/FTS Team!

Post-Launch Sensitivity:

Increased by 4× !!!

special thanks to D. Rigopolou, R. Hopwood, E.

Polehampton


THANKS SPIRE/FTS Team!

Post-Launch Sensitivity:

Increased by 4× !!!

special thanks to D. Rigopolou, R. Hopwood, E.

Polehampton


Key science themesResolved Gas:

CO Excitation Conditions

αCO & SF-law

XDR vs. PDR

Ionized tracer

Diffuse Gas Fraction

Dust: careful test of emissivity slope/flattening.

High-Redshift prescriptions

CO 4-3CO 5-4CO 6-5CO 7-6

CO 8-7CO 9-8CO 10-9[NII] h205

[CI] h609[CI] h369

Rigopoulou+


Ground-CO

Koda%et%al.%2009%

CO(150)(

CARMA%+%%Nobeyama%45m%

CARMA:%1517poindng%mosaic%Nobeyama:%OTF%mapping%1%sigma%~%1x105%Msun;%typical%GMC%~%4sigma%

Excellent(UV(coverage(• %CARMA:%15%antennas%• %Synthesis%observadons%• %Single7dish%coverage%

CARMA%uv%coverage%

Nobeyama%uv%

Sensi>vity(match(%

Nobeyama%

CARMA%(10m+6m)%

Koda%et%al.%2011%

Heracles: The H.E.R.A. CO-line Extragalactic Survey�Stars form out of clouds of molecular gas. The formation and evolution of this star-forming gas represents an important step in building galaxies. We carried out the Heterodyne Receiver Array CO-Line Extragalactic Survey (HERACLES) to constrain how molecular material assembles and forms stars in the nearby universe. Using the powerful Heterodyne Receiver Array (HERA) on the IRAM 30-m telescope (9 beams, 2 polarizations, 13” resolution, Schuster et al. 2004), HERACLES observed molecular gas via the CO J=2-1 transition in a diverse sample of 48 nearby galaxies. By building on previous surveys at other wavelengths - THINGS (Walter et al. 2008), SINGS (Kennicutt et al. 2003), the GALEX NGS (Gil de Paz et al. 2007), and the LVL (Dale et al. 2009) - we are able to observe molecular gas in context. This allows one to compare molecular gas with recently formed stars, the atomic gas reservoir, abundances of dust and heavy elements, kinematics, old stars, spiral arms, and other potentially important drivers for the formation of stars and star-forming clouds.

Team: A. K. Leroy (NRAO, co-P.I.), F. Walter (MPIA, co-P.I.), F. Bigiel (U.C. Berkeley), E. Brinks (Hertfordshire), W.J.G. de Blok (Cape Town), D. Calzetti (U. Mass.), G. Dumas (MPIA), K. Foyle (MPIA), R. Kennicutt(Cambridge), S. Meidt (MPIA), C. Kramer (IRAM), H.-W. Rix (MPIA), E. Schinnerer (MPIA), K. Sandstrom (MPIA), A. Schruba (MPIA), K. Schuster (IRAM), A. Usero (OAN, Spain), A. Weiss (MPIfR), H. Wiesemeyer (MPIfR); the IRAM 30-m is operated by IRAM, which is supported by INSU/CNRS (France), MPG (Germany) and IGN (Spain).

Each panel shows molecular gas in context for one of our targets, with targets approximately in order of stellar mass from top to bottom. From left to right panels show:

Molecular Gas Peak CO intensity From HERACLES

Atomic Gas Column from VLA 21cm data

THINGS + new & archival

Kinematics Here from HI line

Also available from CO

Old Stars Near infrared intensity From SINGS and LVL

Recent Star Formation Composite of FUV (GALEX), mid-IR

(SINGS/LVL), and Hα (SINGS/LVL)

NGC 4579 NGC 4569 NGC 7331

NGC 4254 NGC 4594

NGC 5055 NGC 5194

NGC 5457 NGC 2841 NGC 4321

NGC 3627 NGC 2903 NGC 5713 NGC 3521

NGC 4736

NGC 4536

NGC 2798

NGC 4214

NGC 3184

NGC 3351

NGC 5713

NGC 4236

NGC 4725 NGC 3190

NGC 0337

NGC 3077

NGC 4631 NGC 6946

NGC 3198

NGC 0925

NGC 0628

NGC 3049

NGC 4625

NGC 2976

NGC 2366

IC 2574

NGC 2403

NGC 5474

Holmberg II

Holmberg I DDO 154

M81 Dwarf B

M81 Dwarf A DDO 053

DDO 165

NGC 2146

NGC 3034

Canon (1⇒0): Koda+

The LCO(3!2)–LFIR correlation in the SINGS sample 27

Figure C10. CO J=3-2 images for NGC 3521. (a) CO J=3-2 integrated intensity image. Contours levels are (0.5, 1, 2, 4, 8) K km s!1

(TMB). (b) CO J=3-2 overlaid on a Digitized Sky Survey image. (c) Velocity field. Contour levels are (558, 613, 668, 723, 778, 833, 888,943, 998) km s!1. (d) The velocity dispersion !v as traced by the CO J=3-2 second moment map. Contour levels are (4, 8, 16, 32) kms!1. Similar images derived from the same data have been published in Warren et al. (2010).

c" 0000 RAS, MNRAS 000, 000–000

NGLS (3⇒2): Wilson+

Heracles (2⇒1): Leroy/Walter+

Carma Sting (1⇒0): Bolatto


Example: NGC3627

Log 1

0([O

III]

/Hβ)

log ([NII]/H_)

log ([OIII]/H

`)10 Starforming

galaxies

Seyferts

LINERs

Log10([N II]/Hα)

Nuc

Bar S

Bar N


Example: NGC3627

0

1

2

3

4

5

200 300 400 500 600

Flux (10-18 W m-2 Hz-1 sr-1)

Observed Wavelength (µm)

Nuc

N-Bar

S-Bar

4-35-46-5

7-68-7

9-810-9

11-10

[CI][CI]

[NII]

H2O

+


Example: NGC3627

0

5

10

15

20

25

30

35

1 2 3 4 5 6 7 8 9 10 11 12

12CO I(J) / I(J=0)

J

NGC3627Nuc

N-BarS-Bar


Mapping: [NII] 205µmNGC3521


Mapping: [NII] 205µmNGC3521


[NII] 205µm MappingNGC3627






[NII 205µm]NGC4569


How do Galaxies Assemble Stars?

Kennicutt 1997

The Story, 15

years ago


How do Galaxies Assemble Stars?

J. Freundlich et al.: Towards a resolved Kennicutt-Schmidt law at high redshift

Fig. 2. [OII] line and CO luminosities (respectively left and right panels for each galaxy) in position-velocity planes correspondingto the DEEP2 slits. Smoothed ensembles of clumps are separated by eye along the vertical axis, as shown with the white horizontallines. The [OII] diagrams were normalized in order to have fluxes proportional to the SFR, but all galaxies share the same arbitrarycolor scales. One arcsecond corresponds approximately to 8.5 kpc (Table 1).

Fig. 4. Kennicutt-Schmidt diagram for 1” ensembles of clumpsof EGS13003805 (green), EGS13004291 (red), EGS12007881(blue), and EGS13019128 (purple), using the Kewley et al.(2004) [OII] SFR calibration. The dotted diagonal lines corre-spond to constant gas depletion times of 0.1, 1, and 10 Gyr fromtop to bottom, and the solid black line to a constant depletiontime equal to the mean depletion time of the clumps, tdepl=1.9Gyr. The error bars of 0.3 dex yield a reduced �2 close to 1 andcorrespond to a factor 2 uncertainty, which is a lower estimate.The gray data points from Genzel et al. (2010) and Tacconi et al.(2013) are indicated for comparison with whole galaxies.

4. Discussion and conclusion

4.1. Advantages of the method

We have shown how the various ensembles of clumps can beseparated by their kinematics in PV diagrams, even though theangular resolution of molecular gas and SFR data was not ableto separate them in integrated intensity. As such, the KS diagramcan be obtained within regions of the resolution size (⇠1”).

Previous resolved KS work at high redshift was only ob-tained using serendipitous amplification by gravitational lenses.Decarli et al. (2012) carried out one of the first spatially re-solved studies in high redshift galaxies, using [NII], FIR, andCO observations for two gravitationally lensed z⇠3.9 galaxies.They obtain a steep relation of slope N = 1.4 ±0.2 between thedust continuum and the molecular gas surface brightness. Stronglenses are rare, and determination of the clumps physical param-eters depend on the lensing model. Our method is probably moreappropriate for a systematic study of the star formation at highredshift, until higher resolution instruments resolve the clumps.

In the absence of high resolution molecular gas data,Swinbank et al. (2012) report adaptive optics H↵ observations ofeleven kpc-scale star forming regions identified in z = 0.84�2.23galaxies and measure the velocity dispersion of the ionized gas�and the star formation surface density ⌃SFR. By assuming that �also corresponds to the dispersion of the cold clumps and that theclumps are marginally stable with a Toomre parameter Q ' 1,they claim a correlation between the gas surface density ⌃gas and⌃SFR. But the method is highly indirect, relies on many assump-tions, and underestimates beam-smearing e↵ects in the determi-nation of �.

5

Freundlich+ 2013

2 Federrath

Figure 1. Star formation rate column density (⌃SFR) versusgas column density (⌃gas), measured in MW clouds, as well asin nearby and high-redshift disc and starburst galaxies. The datashown are from W10 (Wu et al. 2010): HCN(1–0) clumps (upar-rows); from H10 (Heiderman et al. 2010): Taurus (filled square),class I YSOs and flat SED YSOs (green and red stars with up-per limits shown as downarrows), and C2D+GB clouds (opensquares); from L10 (Lada et al. 2010): molecular clouds observedat two di↵erent extinction thresholds (AK > 0.1mag: open circles,and AK > 0.8mag: filled circles); from G11 (Gutermuth et al.2011): class II YSO counts in eight molecular clouds (crosses);from Y09 (Yusef-Zadeh et al. 2009): the Central Molecular Zone(CMZ, turquoise diamond with error bars); and from B11 (Bo-latto et al. 2011): the Small Magellanic Cloud (SMC, red trian-gle with error bars). Extragalactic data (Kennicutt 1998; Boucheet al. 2007; Daddi et al. 2010; Genzel et al. 2010; Tacconi et al.2010) of disc (D) and starburst (SB) galaxies at low redshift(z = 0) and high redshift (z ⇠ 1–3) are reproduced from the tab-ulated compilation in KDM12 (Krumholz et al. 2012). (Table 3in KDM12 for Ds and SBs contains naming errors and SB galaxyVII Zw 31, erroneously called ‘NGC 7552’, has wrong ⌃gas and⌃SFR. An erratum is in preparation [M. Krumholz, private com-munication] and those errors have been corrected here.) Typicaluncertainties for the Ds and SBs are of the order of 0.5 dex (fac-tor of 3) in both ⌃gas and ⌃SFR (Kennicutt 1998), but there maybe additional uncertainties due to calibration errors caused bydi↵erent forms of the initial mass function and di↵erent CO/H2

conversion factors (Daddi et al. 2010). Previously suggested starformation laws from extragalactic observations by K98 (Kenni-cutt 1998) and B08 (Bigiel et al. 2008), as well as from MWobservations by W10 and H10 are shown as lines for comparison.

2 A MORE UNIVERSAL STAR FORMATIONLAW

More recently, Krumholz et al. (2012, hereafter KDM12)thus argued that the standard star formation relation shownin Figure 1 may not provide the best physical representa-tion. Based on the assumption that the SFR is inversely pro-portional to the dynamical time of the gas (Schmidt 1959;Elmegreen 2002), KDM12 suggest that a better fit is ob-tained, if ⌃SFR is plotted against ⌃gas/t↵ , i.e. ⌃SFR as afunction of ⌃gas divided by the local gas collapse time,

t↵(⇢) =

✓3⇡

32G⇢

◆1/2

, (1)

evaluated for each cloud or galactic system individ-ually. Although not directly observable, the gas density⇢ = (3

p⇡/4)M/A3/2 with the cloud mass M = ⌃gasA and

the observed area A can be estimated by assuming that theclouds are approximately spherical objects (KDM12), intro-ducing additional uncertainties (Appendix A). For extra-galactic systems, the gas collapse time is taken to be theminimum of the Toomre time for stability of the disc orstarburst and the local cloud freefall time (see KDM12 fordetails1). In this way, the MW clouds and the extragalacticdata seem to exhibit a much tighter correlation, which isshown in Figure 2(a). KDM12 only included the C2D+GBclouds from H10 and the L10 clouds at the two di↵erentextinction thresholds, while here we add all data from H10,Wu et al. (2010, hereafter W10), and the clouds observedin Gutermuth et al. (2011, hereafter G11). We also includean average of the 200 pc resolution data (A = 4.5⇥ 104 pc2;A. Bolatto, private communication) of the Small MagellanicCloud (SMC) (Bolatto et al. 2011). One might questionwhether mixing resolved measurements of MW clouds andgalactic discs with unresolved discs and starbursts (KDM12)in a single plot produces a physically meaningful compari-son, because of extinction and telescope resolution issues(e.g. Calzetti et al. 2012; Shetty et al. 2013). Encouragingly,however, we find in tests with synthetic observations at dif-ferent extinction thresholds and telescope resolutions vary-ing by a factor of 32 that measurements presented in theform of Figure 2 vary by less than a factor of two for fixedphysical conditions (Appendix B).

The dashed line in Figure 2(a) shows the empirical re-lation by KDM12,

⌃SFR = ✏SF,0 ⇥ ⌃gas/t↵ , (2)

with a constant proportionality factor, ✏SF,0 = 1%, whichwe define here as the total star formation e�ciency, ✏SF,0 ⌘✏⇥SFE. In this expression for ✏SF,0, the local core-to-star ef-ficiency, ✏ = 0.3–0.7, is the fraction of infalling gas that is ac-creted by the star (Matzner & McKee 2000), i.e. about half.The other half is expelled by jets, winds and outflows. Theglobal (cloud-scale) e�ciency, SFE = 1%–6%, is the typi-cal fraction of gas forming stars in a whole molecular cloud(Evans et al. 2009; Lada et al. 2010; Federrath & Klessen2013). This yields a combined, total star formation e�ciency,✏SF,0 ⇠ 0.3%–4.2%. Here we adopt an intermediate value,✏SF,0 = 1%, as favoured in observations and analytic models(Krumholz & Tan 2007; Renaud et al. 2012); however, wealso study the influence of varying ✏SF,0 below. The observa-tional data in Figure 2(a) indeed exhibit a better correlationthan in Figure 1, yet the scatter is still significant and re-mained largely unexplained in KDM12. What is the originof this persistent scatter?

To advance on this issue, we compare the observa-tions with computer simulations from Federrath & Klessen(2012, hereafter FK12), covering a substantial range ofobserved physical cloud parameters with Mach numbersM = �v/cs = 5–50, di↵erent driving of the turbulenceparametrized by b = 1/3 for solenoidal (divergence-free),b = 0.4 for mixed and b = 1 for compressive (rotation-free)driving, as well as a few di↵erent magnetic field strengths

1 The high-z disc and starburst galaxy dataset in table 4 ofKDM12 contains errors related to the computation of the Toomretime [M. Krumholz, private communication]. Figure 2 here showsthe corrected data.

c� 0000 RAS, MNRAS 000, 000–000

Federath, 2013

The Story, today


What we want: The H2 molecule

most abundant molecule by 104×



×most abundant molecule by 104×



×most abundant molecule by 104×

Not Useful for

Studying “bulk”

Molecular Gas


What we get: The CO Molecule

50k



50k



XCO ' 4.4 M�/pc2/K km s�1

50k


CO “SLED”30 Weiss et al.

Figure 3. Normalized CO line SEDs for all sources observed so far in oursurvey. The SEDs are normalized by the H2 mass (CO(1–0) flux density).We also added the QSO SDSS J1148+5251 at redshift z = 6.4 (Walter et al.2003; Bertoldi et al. 2003) and BR1202!0725 at z = 4.7 (Riechers et al.2006, and this volume) to this plot. The SEDs of the QSOs are shown assolid lines, those for the SMGs as dashed lines. The decreasing order of thesource names corresponds to the decreasing CO excitation as traced by theirCO SEDs. The CO line SED of SMMJ04431+0210 is still poorly constrained,but its low CO(5–4)/CO(3–2) ratio indicates that it has excitation similar tothat of SMMJ16359+6612 . All four submm galaxies show low CO excitationcompared to the QSOs in our sample.

the CO flux comes from an “overlap” region between two merging galaxies (seeWeiß et al. 2005a). This geometry has been suggested for SMM J16359+6612and ERO J16450+4626 (Kneib et al. 2004; Greve et al. 2003). For compari-son, we therefore also show the CO SED for the overlap region in the Antennaein Figure 4 (Zhu, Seaquist, & Kuno 2003). Although no observations abovethe CO(3–2) line have been published in the literature so far, the low averageCO(3–2)/CO(1–0) line ratio suggests that the gas in the Antennae overlap re-gion has a lower excitation than starburst galaxies/QSOs and is more similar

Weiss+ 2007Thursday, October 17, 13

High-z applications314 Astrophys Space Sci (2008) 313: 313–316

Fig. 1 ALMA CO ‘discovery space’: the horizontal lines indicatewhich CO transition (plotted on the y-axis) can be observed with whichALMA band as a function of redshift (plotted on the x-axis). For ob-jects with z > 7, only the higher-J CO transitions can be observed withALMA. The [CII] ‘discovery space’ is also indicated

graphically illustrated in Fig. 1 where we plot ALMA’s ‘COdiscovery space’ (i.e., which line can be observed at whichredshift using which ALMA band). The high-J transitionscorrespond to highly excited gas (either due to high kinetictemperatures, high densities, or both) which may not be ex-cited in normal starforming environments. This is shown inFig. 2 (taken from Weiss et al. 2007) where measured COline strengths (as a function of J , this is sometimes referredto as CO line ladders/SEDs) are plotted for a number ofkey sources. What is immediately obvious from this plot isthat most objects have sharply decreasing CO line strengthsbeyond J > 8, in particular starforming systems such asNGC 253, or the sub-millimeter galaxy plotted in this dia-gram (the quasars appear to be more excited, but their COline SED still turns over at J ! 7, for an exceptional objectsee APM 07279, Weiss et al. 2007). This comparison im-mediately implies that emission from the CO molecule willtypically be very difficult to observe with ALMA at z > 7 asthe observable lines will simply not be excited.

2.2 Ionized carbon ([CII]) to the rescue!

An alternative tracer of the interstellar medium is one ofthe main cooling line of the ISM, the 2P3/2 " 2P1/2 fine-structure line of C+ (or [CII]). In brief, the [CII] line isexpected to be much stronger than any of the CO lines.Given its high frequency (157.74 µm, corresponding to1900.54 GHz) [CII] studies in the local universe are lim-ited to airborne or satellite missions (e.g. Stacey et al. 1991;Malhotra et al. 1997; Madden et al. 1997). These studies

Fig. 2 Comparison of various normalized (by their CO(1–0) flux den-sity) CO line SEDs at low and high redshift (figure taken from Weisset al. 2007). The CO line SEDs decline rapidly beyond J = 6–8

have demonstrated that this single line can indeed carry agood fraction of the total infrared luminosity (LFIR) of anentire galaxy. In the local universe, the ratio LCII/LFIR hasbeen found to be 2–5 # 10$4 in the case of ULIRGS (e.g.Gerin and Phillips 2000), but is more like 5–10 # 10$3 inmore typical starforming galaxies (for a discussion on pos-sible reasons for the suppressed ratio in ULIRGs see, e.g.,Luhman et al. 1998). Notably, the ratio has been found to be1% or even higher in low metallicity environments. E.g., inthe low-metallicity galaxy IC10, LCII/LFIR reaches valuesas high as 4%, with an average value of 2% (Madden et al.1997; see Israel et al. 1996 for a similar result for the LMC).This is the reason why it has long been argued (e.g., Stark1997) that observation of the [CII] line of pristine systemsat the highest redshifts will likely be the key to study mole-cular gas in the earliest starforming systems, in particularin the era of ALMA. The ALMA [CII] ‘discovery space’ isalso indicated in Fig. 1.

3 Expected [CII] line strengths

At the redshifts of the LAEs, the [CII] line is shifted to the1 mm band of ALMA (band 6, 211–275 GHz). [CII] emis-sion has recently been successfully detected using the IRAM30 m in the highest redshift quasar J1148+5251 at z = 6.42(Maiolino et al. 2005, see Fig. 3). The notable differencebetween J1148+5251 and the z > 6 LAEs is that the ratioLCII/LFIR has been found to be very low (!5 # 10$4) in

Walter & Carrili, 2008

“The ability to detect spectral line emission

from CO or CII in a normal galaxy like the Milky Way at a redshift of z = 3, in less than 24 hours of

observation.”

12/26 have two or more lines detected.J. Vieira


Use ALMA?

✓ Fantastic Sensitivity.

× Very limited Field of View.

ALMA is not ideally suited to providing a calibration of gas excitation.

ALMA CO 6-5 FOV


Use ALMA?

✓ Fantastic Sensitivity.

× Very limited Field of View.

ALMA is not ideally suited to providing a calibration of gas excitation.

ALMA CO 6-5 FOV


Sandstrom+ 2013

The Astrophysical Journal, 777:5 (33pp), 2013 November 1 Sandstrom et al.

(a) (b)

(c) (d)

Figure 1. Illustration of the technique to determine the DGR and !CO for a single solution pixel in NGC 6946. Panel (a) shows a portion of the CO map fromHERACLES overlaid with the half-beam spaced sampling grid. The solution pixel in question is shown with a solid white hexagon and the 37 individual samplingpoints included in the solution are shown in red. Two neighboring solution pixels are also highlighted with a dashed white line to show how the pixels are arrangedand that neighboring solution pixels share !40% of their sampling points. Panel (b) illustrates that the scatter in the DGR changes as a function of !CO. Here, we haveplotted histograms of the measured DGRi values in this solution pixel at three different values of !CO (the optimal value in black and a factor of five above and belowthis value in green and purple, respectively). Panel (c) illustrates that this scatter originates from the variation of the DGR as a function of CO/H i when !CO is not atthe optimal value. We show this effect by plotting the 37 DGRi values as a function of CO/H i ratio for the same three !CO values shown in panel (b). The horizontallines indicate the mean DGR for each set of points. The slope in the DGR vs. CO/H i space is minimized at the optimal !CO. Finally, in panel (d), we show the scatterin the DGR at each value of the full !CO grid shown in black. The three highlighted !CO values are marked with vertical lines. Panel (d) highlights the fact that theDGR scatter is minimized at the best-fit log(!CO) = 0.15 ± 0.22 in this region. The minimization of the scatter in log(DGR), as shown in panel (d), is the techniquewe have determined to be the most effective for determining !CO and the DGR, using tests that are described in detail in the Appendix.(A color version of this figure is available in the online journal.)

here, we use an !CO grid with 0.05 dex spacing, spanning therange !CO = 0.1–100 M" pc#2 (K km s#1)#1.

We determine the most “uniform” DGR in a solution pixelby minimizing the scatter in the DGRi values as a functionof !CO. The scatter is measured with a robust estimator of thestandard deviation37 of the logarithm of the DGRi values—thistechnique appears to work best because outliers have little effecton the measured scatter because of both the logarithmic unitsand the outlier suppression. After measuring the scatter in theDGR at every !CO value, we find the !CO at which the scatter(!log(DGR)) is minimized. This value is taken to be our best-fit!CO value for the region. We consider a solution to be foundwhen a minimum has been located in the DGR scatter withinthe range of our !CO grid. This outcome does not occur inevery solution pixel—some pixels do not have sufficient CO/H icontrast, others have too low S/N, and, for some, the minimumis at the edge of the allowable range and the solution is not well

37 We use the IDL implementation of Tukey’s biweight mean (Press et al.2002) biweight_mean.pro.

constrained. The failed solutions are not included in our furtheranalysis.

An illustration of the technique is shown in Figure 1 for aregion in NGC 6946. In panel (a), we show the HERACLES COJ = (2–1) map of the galaxy with our half-beam sampling grid;the hexagonal region shows the “solution pixel” in question,which includes 37 individual samples from the maps. Panels (b)and (c) show how varying !CO affects the mean DGR and thescatter for the points in the region, illustrating how the scatterincreases away from the best !CO value. Panel (d) shows thescatter as a function of !CO for the whole !CO grid. A clearminimum exists for this solution pixel at !CO ! 1.4 M" pc#2

(K km s#1)#1.

3.3. Statistical Uncertainties on !CO and the DGR

We judge the uncertainties on the “best-fit” !CO and the DGRin several ways. First, to take into account statistical errors, weperform a Monte Carlo test on the solutions by adding randomnoise to our measured "D, "H i, and ICO values according to each

7


Figure 2. H i, CO, and !D maps for NGC 0628, from left to right (D = 7.2 Mpc; 1!! = 35 pc). The centers of the pixels in which we perform the simultaneous !COand DGR solutions are shown as circles overlaid on the images. The gray cross in each panel shows the central solution pixel for the galaxy. In the middle panel, thecoverage of the HERACLES CO map is shown with a dotted line. Similar plots for all galaxies in the sample can be found in the Appendix.(A color version of this figure is available in the online journal.)

If opaque H i exists at the level Braun et al. (2009) found inM31, it would have two main effects on our solutions for !CO.First, on average, the optically thin estimate for the atomic gasmass would be too low, resulting in our procedure determininga DGR that is too high (excess dust compared to the amount ofgas). In the molecular regions, then, we will expect too much gasbased on that same DGR, and consequently artificially increase!CO. Second, since the opaque H i features do not appear to bespatially associated with molecular gas (see Braun et al. 2009,Section 4.1, for a further discussion), these features will act asa source of intrinsic scatter in the DGR. In the Appendix, weexplore the effect of intrinsic scatter on our solution technique.At the level of opaque H i in M31, we do not find an appreciablebias in the recovered !CO due to scatter. We expect the magnitudeof the systematic effects due to opaque H i, if it exists, to bewell within the statistical uncertainties we achieve on the !COmeasurements.

3.5. Systematic Uncertainties on the DGR

3.5.1. Absolute Calibration of !D

As we have discussed above, as long as !D is a linear tracerof the true dust mass surface density within a given solutionpixel, its absolute calibration has no effect on the !CO valuewe measure. The same is not true for the DGR value. Anyuncertainties on the calibration of !D will be directly reflectedin the DGR measurement. The !D values we used are from fitsof the Draine & Li (2007) models to the IR SED using the MWRV = 3.1 grain model. The extent to which the appropriatedust emissivity "# deviates from the value used by this modelrepresents a systematic uncertainty on the DGR values wederive. Our knowledge of "# in different environments is limited,but there are constraints from observations of dust extinctioncurves and depletions in the Large Magellanic Cloud (LMC)and Small Magellanic Cloud (SMC; cf. Weingartner & Draine2001), where measured RV values can deviate significantly fromthe canonical value of 3.1. Draine et al. (2007) demonstrated thatthe !D values decreased by a factor of "1.2 when the LMC orSMC dust model was used instead of the MW RV = 3.1 model.

Given that our sample is largely dominated by spiral galaxiesand hence does not probe environments with metallicitiescomparable to those of the SMC (due to the faintness of COin such regions and our S/N limitations), we expect that thesystematic uncertainties on our DGR when comparing withother results from Draine & Li (2007) model fits is small. Itis important to note, however, that different dust models, evenfit to the same RV = 3.1 extinction curve, have systematicoffsets in their dust mass predictions due to different grain sizedistributions, grain composition, etc. Therefore, the comparisonof our DGR values to results from studies not using the Draine& Li (2007) models will show systematic offsets.

4. RESULTS

4.1. NGC 0628 Results Example

We divided each of the 26 galaxies in our sample into solutionpixels and performed the simultaneous solution for the DGRand !CO in each pixel. As an example, we present the results forNGC 0628 in the following section. The results for all solutionpixels in all galaxies can be found in the Appendix.

Figure 2 shows, from left to right, the H i, CO, and !D mapsused in our analysis. The circles overlaid on the maps representthe centers of the solution pixels we have defined. Figure 3shows the same circles representing the solution pixel centers.The left panel shows the pixel centers now filled in with a colorrepresenting the best !CO solution for that pixel. In the middlepanel, a gray scale shows the uncertainty on that !CO solution.The DGR values are shown in the panel on the right. Finally, inFigure 4, we show these measured !CO values as a function ofgalactocentric radius (r25). For comparison, Figure 4 also showsthe local MW !CO = 4.4 M# pc$2 (K km s$1)$1 value with asolid horizontal line (dotted lines show a factor of two aboveand below; see Section 5.1 for details on the measurement ofthe MW value). We note that the MW may show a gradient of!CO with radius (also discussed in Section 5.1), but for purposesof comparison with the most widely used conversion factor, weuse a constant !CO on all plots.

9


Figure 15. !CO as a function of metallicity compared to previous measurements and models. The left panel shows measurements with the PT05 calibration while theright panel shows measurements with the KK04 calibration. Measurements from this paper are shown as gray points (individual solutions) and red circles (galaxyaverages). We show all of the solution pixels where ICO > 1 km s!1 and i > 65", regardless of the source of the metallicity measurement (i.e., without the requirementof having measured H ii region metallicities from M10). Measurements of !CO in Local Group galaxies from Leroy et al. (2011) are shown with green circles. TheMW !CO is shown with a gray line and the average of our solution pixels with no weighting is shown with a dotted black line. Predictions based on the model ofWolfire et al. (2010) are shown with a purple dot-dashed line and those based on Glover & Mac Low (2011) are shown with a dashed blue line. These predictionsassume a linear dependence of the DGR on metallicity and a fixed gas mass surface density for molecular clouds of !GMC = 100 M# pc!2. The model predictionsare normalized to have !CO = 4.4 M# pc!2 (K km s!1)!1 at the metallicity adopted for the Milky Way in each calibration. Note that the metallicities we use forNGC 5457 are not from strong-line calibrations, so they appear in the same position in both plots. Regardless of the metallicity calibration, our measurements do notextend to low enough metallicities to constrain the effects of “CO-dark” H2.(A color version of this figure is available in the online journal.)

In a $kiloparsec region of a galaxy, many individual molecu-lar clouds will be averaged together in our measurement. There-fore, another factor that could contribute to the variation of ourmeasured !CO is changes in the cloud populations. In addition,any significant component of diffuse CO emission would be in-cluded as well. Several studies have suggested that CO emissionfrom diffuse gas may not be negligible, even in the local area ofthe MW (Liszt & Lucas 1998; Goldsmith et al. 2008). Interest-ingly, it appears that locally, !CO is similar in diffuse moleculargas and self-gravitating GMCs (Liszt et al. 2010). Recent workby Liszt & Pety (2012) has shown that under differing environ-mental conditions, the conversion factor appropriate for diffusegas can vary substantially. Therefore, depending on the amountof diffuse molecular gas and the local conditions in a galaxy,!CO may be very different from what is observed in the localarea of the MW. The contribution of diffuse molecular gas witha different !CO has been cited previously as a cause for discrep-ancies between different techniques for measuring !CO in M51(Schinnerer et al. 2010).

To summarize our theoretical expectations, !CO measuredin $kiloparsec regions of nearby galaxies can vary as a func-tion of environment for many reasons: changes in the excitationtemperature and density of the gas or contributions from pres-sure or magnetic support in self-gravitating clouds; envelopesof “CO-dark” gas; changes in the molecular cloud population;or contributions from diffuse CO emission. Unfortunately, di-rectly measuring temperature, density, and velocity dispersionin large samples of extragalactic GMCs is challenging due tothe need to resolve individual clouds in multiple molecular gasemission lines. Therefore, we are left with more indirect tracersof changes to the GMC properties. In Section 4.6, we examinedthe correlation of !CO with U , qPAH, metallicity, !%, !SFR, !D,and galactocentric radius.

The average radiation field U , measured from the dust SED,could influence !CO through the gas excitation temperature.

If photoelectric heating dominates over other heat sources (e.g.,cosmic rays) at the "CO $ 1 surface, then the intensity of theradiation field may play a role in determining Tex (Wolfire et al.1993). Likewise, qPAH and metallicity could both influence theefficiency of the photoelectric effect (Bakes & Tielens 1994;Rollig et al. 2006), thereby changing the heating rate. We expectthat !SFR should be responsible for higher U in many regions.Enhanced SFRs leading to higher radiation field intensitieshas been suggested as the explanation for the observed high!CO in several outer-disk molecular clouds in M33 (Bigielet al. 2010). If the gas excitation temperature were affectedby the radiation field or photoelectric heating, we would expectnegative correlations (i.e., lower !CO at higher U , !SFR, qPAH,etc.). The correlations we observe between !CO and U , qPAH,and !SFR are generally weak. U and !CO show the strongestassociation and the slope of the trend is negative, with lower!CO at higher U . This result is consistent with the expectationof higher radiation field intensities leading to warmer moleculargas temperatures, but the correlation is weak and other variablesmay play a more important role.

We see a weak trend of !CO with metallicity. Since ourobservations are limited to regions with metallicities similarto or higher than that of the MW, we may not expect to see astrong correlation between !CO and metallicity. The fraction ofgas mass in the layer where CO is photodissociated is predictedto be $30% at the MW metallicity/DGR (Wolfire et al. 2010).Thus, increasing the DGR should only have a minimal effect onthe conversion factor at MW metallicity and above. To illustratethis point, in Figure 15 we plot our !CO measurements as afunction of metallicity and overlay the Leroy et al. (2011) LocalGroup measurements and the models of Wolfire et al. (2010)and Glover & Mac Low (2011). The model predictions arenormalized to !CO = 4.4 M# pc!2 (K km s!1)!1 at the MWmetallicity (i.e., 12 + log(O/H) = 8.5 in PT05 and 8.8 in KK04;from the Orion Nebula). Both models assume a linear scaling of

23

XC

O


NGC 3627

NGC 6946

NGC 3351NGC 4321NGC 4736

NGC 5055

NGC 4254

Evidence for enhanced CO excitation in centers with

low αCO from BTP.

Log(αCO)

BTP Pointings towards galaxy centers

Many galaxy centers peak at J~6, similar to M82.

0.1

1

10

200 300 400 500 600

Flux

(10-1

8 W m

-2 sr

-1)


[NII]

[CI][CI]

0.1

1

10

200 300 400 500 600

Flux

(10-1

8 W m

-2 sr

-1)


4-35-4

6-57-68-79-8

12COJ=(10-9)

[NII][CI]

[CI]

NGC 4254

NGC 6946

Sandstrom+, in prep


Empirical Correlations

What is the typical range of CO excitation conditions?

What 2nd order parameters control <J>?

Resolved BtP SLEDs at 5kpc

scale

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

4 5 6 7

IJ / IJ=4

Jupper

4.6

4.7

4.8

4.9

5.0

5.1

5.2

5.3

Y J×I / Y I (J=4...6)


NGC1266

Buried at the core of a “red and dead” galaxy at 30Mpc

109 M⊙◉☉ molecular Gas disk packed into 60pc: similar size, 100× more mass than a Giant Molecular Cloud.

The end.

questions?"

The background image is g+r of NGC 1266 from MEGACAM (Duc et al., in prep)."

Special thanks to David Meier (NMT), Kristina Nyland (NMT) and Tim Davis (ESO)

⌃H2 ⇠ 3⇥ 104 M�/kpc2

alatalo+ 2011


Pelligrini+, ApJL submitted

A CASE FOR SHOCKED WATER IN NGC 1266 3

0

2

4

6

8

10

12

200 300 400 500 600

Flux

(Jy)


12COJ=(4-3)

5-46-5

7-68-79-810-911-1012-11

13-12

[NII]

H2O

[CI] 0 1 2 3 4 5 6 7 8

255 260 265 270 275 280

H2O+

12CO (10-9)

H2O

Figure 1. The SPIRE-FTS spectrum from 200 to 600 µm. Emission lines include 12CO from J=(4-3) to J=(13-12), [N II] 205, [C I] 370 and 609 µm. Strongortho and para H2O emission characteristic of ULIRGS is seen in emission (solid line) as well as absorption (dashed line).

tion from H2O+, but this has been seen in systems whereXDR are ruled out (e.g. Arp220, Rangwala et al. 2011).

Second, !CO also eliminates XDRs as the dominant heat-ing source in NGC 1266. The efficiency at which X-rays areconverted to line emission and IR continuum is very differ-ent than in shocks which heat gas but are relatively ineffi-cient at heating dust. This is illustrated in Figure 3, a plot of!CO against ![CI] and !H2O. As discussed in Meijerink et al.(2013), XDR and PDR models also place an upper limit onthe line-to-continuum ratio of !CO ! 10!4, represented bythe dashed line in Figure 3. In NGC 1266 that ratio for the 13detected CO lines is nearly 10!3, just below that of NGC 6240(see Figure 3). Alatalo et al. (2011) found that the observed!H2

ratio was 3 times larger than could explained by an XDR.The ratio of ![CI] is also enhanced by an order of magnitudein NGC 1266, as in NGC 6240. The linearity of [CI] with12CO, seen in Figure 3, suggests it is a powerful diagnosticespecially useful where the entire 12CO is not observable, asis the case with high redshift galaxies observed with ALMA.

3.2. Shocked H2O

Table 1 includes emission from H2O, a molecule whose gasphase abundance is driven by two processes: formation viaendothermic reactions and the release from ices on the surfaceof dust grains. J- and C-shocks, as well as XDRs, are capableof producing the warm and dense molecular gas characteristicof H2O emitting regions.

Typically H2O emission is very faint in PDRs alone.Since individual H2O line emission is not included in theWolfire et al. (2010) PDR models we assume the Orion Barvalue of L(H2O)/L(CO)= 0.001 (Habart et al. 2010) for ourmodeled PDRs. The observed parameters of the Orion BarPDR G0 " 4#104 (Tielens & Hollenbach 1985) and n(H) "105 cm!3 (Allers et al. 2005; Pellegrini et al. 2009) are ofhigher density and excitation than the PDR parameters nec-essary to explain the low excitation CO in NGC 1266. Theseenergetic conditions favor the formation of water, and thusprovide an upper limit to the H2O emission we expect fromthe PDR.

Table 2 and Figure3 summarize the ratio of H2O, 12CO and

Table 1NGC 1266 IRS, PACS and SPIRE-FTS Extinction Corrected Line

Fluxes

Species Trans Wave(µm) Flux(10!17 W m!2) Inst.

12CO 1-0 2600 0.063±0.004 1

12CO 2-1 1300 0.39±0.015 1

12CO 3-2 867.0 1.02±0.11 1

12CO 4-3 650.7 2.69±1.14 2

12CO 5-4 520.6 3.84±0.69 2

12CO 6-5 433.9 4.28±0.35 2

12CO 7-6 371.9 4.76±0.28 2

12CO 8-7 325.5 4.71±0.36 2

12CO 9-8 289.3 4.64±0.65 2

12CO 10-9 260.4 3.94±0.60 2

12CO 11-10 236.8 4.03±0.61 2

12CO 12-11 217.1 2.85±0.30 2

12CO 13-12 200.4 2.65±0.58 2

H2O 211 ! 202 398.9 1.09±0.32 2

H2O 202 ! 111 303.7 2.91±1.04 2

H2O 312 ! 303 273.4 1.80±0.40 2

H2O 312 ! 221 260.2 1.83±0.61 2

H2O 321 ! 312 258.0 2.38±0.45 2

H2O 220 ! 211 244.1 1.69±0.45 2

H2O 422 ! 413 248.2 <0.47 2

H2O 523 ! 514 212.5 <0.47 2

[C I] · · · 370.7 3.19±0.26 2

[C I] · · · 609.6 1.13±0.86 2

[C II] · · · 157.7 17.30±0.46 3

[O I] · · · 63.18 11.40±1.32 3

[N II] · · · 205 1.70±0.88 2

H2 S(4) 8.026 21.90±1.01 4

H2 S(3) 9.662 13.00±0.39 4

H2 S(2) 12.282 10.90±0.26 4

H2 S(1) 17.035 8.48±0.10 4

H2 S(0) 28.171 2.94±0.22 4

H2 1-0S(1) 2.1213 10.90±0.20 5

Note. — Ground 12CO from 1Alatalo et al. (2011). Lines observedwith 2SPIRE-FTS, 3PACS, 4Spitzer-IRS instruments. NIR H2 from5Riffel et al. (2013). IRS fluxes have been extracted over a 10” aperturecentered on the peak of the IR continuum emission which is unresolvedby Spitzer and Herschel.

NGC1266 Beyond the Peak


v~200km/s outflow @3×SFR

4 PELLEGRINI ET AL.

10-3

10-2

10-1

PDR+ShockPDR

ShockFTS

Ground

10-3

10-2

10-1

�1 �2 �3 �4 �5 �6 �7 �8 �9 �10 �11 �12 �13 �14

Jup

L(1

2C

O�(

J up)�

)/Σ

J�L

(12C

O)

FTSGroundArp220

M82Mrk231

NGC6240Orion�Bar

IC342

0.000.050.100.150.200.250.30

�1 �2 �3 �4 �5 �6 �7

Log(n)�cm-3

Pro

bab

ilit

y�D

istr

ibu

tio

n

0.000.010.020.030.040.050.06

-1 �0 �1 �2 �3 �4 �5 �6 �7

log(G0)

0.000.100.200.300.400.500.600.70

�10 �20 �30 �40

Shock�velocity�(km�s-1

)

Figure 2. (Top-Left) The NGC 1266 12CO SLED from J=(1-0) to (13-12) (circles) normalized by the total observed 12CO luminosity. Jupper from 1 to 3 arefrom ground based observations. Jupper = 4 and above were observed with SPIRE-FTS. Shown is the best fitting PDR (dashed blue) and Shock model (solid

red). (Bottom-Left) Comparison with published 12CO SLEDs of various galaxies, and the Orion Bar. With exception to IC342, bright water emission is detectedin each galaxy, attributed to either an XDR or mechanical heating. (Right) Normalized probability distributions functions of the PDR+Shock parameters. Avertical dashed line represents the only pre-shock density that fits observations.

-5.0

-4.5

-4.0

-3.5

-3.0

-6.0 -5.5 -5.0 -4.5

Log

10�Σ

�L(1

2C

O)/

L(T

IR)

Log10�Σ�L([C�I])/L(TIR)

XDR�and�PDR

NGC1266

Mrk231

NGC6240

Arp220M82

IC342

-7.0 -6.0 -5.0 -4.0 -3.0

Log10�Σ�L(H2O)/L(TIR)

NGC1266

Mrk231

NGC6240

Arp220M82

IC342

Figure 3. Left: Total CO, from J=(1-0) to (13-12), relative to TIR, and [CI]370,609µm relative to TIR for galaxies with detected H2O emission and a detailedanalysis of gas excitation mechanisms. Galaxies are color coded as PDRs (red), XDRs (green), mechanical (teal) and shocks (blue). Right: Same as left, with totalL(H2O)/TIR on the x-axis. The horizontal dashed line marks the upper limit of L(CO)/L(TIR) for XDRs and PDRs (Meijerink & Spaans 2005; Meijerink et al.2007).

Pellegrini+ submitted


Water: The Shocking Truth

0.001

0.01

0.1

1

100 200 300 400 500 600

Temperature (K)

Normalized IH2O

NGC 1266Mrk231

Shock Model


0.000.050.100.150.200.250.30

1 2 3 4 5 6 7

Log(n) cm-3

Probability Distribution

0.000.010.020.030.040.050.06

-1 0 1 2 3 4 5 6 7log(G0)

0.000.100.200.300.400.500.600.70

10 20 30 40

Shock velocity (km s-1)


A Under-Luminous ULIRG?

4 PELLEGRINI ET AL.

10-3

10-2

10-1

PDR+ShockPDR

ShockFTS

Ground

10-3

10-2

10-1

�1 �2 �3 �4 �5 �6 �7 �8 �9 �10 �11 �12 �13 �14

Jup

L(1

2C

O�(

J up)�

)/Σ

J�L

(12C

O)

FTSGroundArp220

M82Mrk231

NGC6240Orion�Bar

IC342

0.000.050.100.150.200.250.30

�1 �2 �3 �4 �5 �6 �7

Log(n)�cm-3

Pro

bab

ilit

y�D

istr

ibuti

on

0.000.010.020.030.040.050.06

-1 �0 �1 �2 �3 �4 �5 �6 �7

log(G0)

0.000.100.200.300.400.500.600.70

�10 �20 �30 �40


)



-5.0

-4.5

-4.0

-3.5

-3.0

-6.0 -5.5 -5.0 -4.5

Lo

g1

0�Σ

�L(1

2C

O)/

L(T

IR)


XDR�and�PDR

NGC1266

Mrk231

NGC6240

Arp220M82

IC342

-7.0 -6.0 -5.0 -4.0 -3.0


NGC1266

Mrk231

NGC6240

Arp220M82

IC342



and.... The significance of line/continuum


Summary

4 PELLEGRINI ET AL.

10-3

10-2

10-1

PDR+ShockPDR

ShockFTS

Ground

10-3

10-2

10-1

�1 �2 �3 �4 �5 �6 �7 �8 �9 �10 �11 �12 �13 �14

Jup

L(1

2C

O�(

J up)�

)/Σ

J�L

(12C

O)

FTSGroundArp220

M82Mrk231

NGC6240Orion�Bar

IC342

0.000.050.100.150.200.250.30

�1 �2 �3 �4 �5 �6 �7

Log(n)�cm-3

Pro

bab

ilit

y�D

istr

ibuti

on

0.000.010.020.030.040.050.06

-1 �0 �1 �2 �3 �4 �5 �6 �7

log(G0)

0.000.100.200.300.400.500.600.70

�10 �20 �30 �40


)



-5.0

-4.5

-4.0

-3.5

-3.0

-6.0 -5.5 -5.0 -4.5

Log

10�Σ

�L(1

2C

O)/

L(T

IR)


XDR�and�PDR

NGC1266

Mrk231

NGC6240

Arp220M82

IC342

-7.0 -6.0 -5.0 -4.0 -3.0


NGC1266

Mrk231

NGC6240

Arp220M82

IC342


0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

4 5 6 7

IJ / IJ=4

Jupper

4.6

4.7

4.8

4.9

5.0

5.1

5.2

5.3

Y J×I / Y I (J=4...6)

[NII] 205µm

314 Astrophys Space Sci (2008) 313: 313–316

Fig. 1 ALMA CO ‘discovery space’: the horizontal lines indicatewhich CO transition (plotted on the y-axis) can be observed with whichALMA band as a function of redshift (plotted on the x-axis). For ob-jects with z > 7, only the higher-J CO transitions can be observed withALMA. The [CII] ‘discovery space’ is also indicated

graphically illustrated in Fig. 1 where we plot ALMA’s ‘COdiscovery space’ (i.e., which line can be observed at whichredshift using which ALMA band). The high-J transitionscorrespond to highly excited gas (either due to high kinetictemperatures, high densities, or both) which may not be ex-cited in normal starforming environments. This is shown inFig. 2 (taken from Weiss et al. 2007) where measured COline strengths (as a function of J , this is sometimes referredto as CO line ladders/SEDs) are plotted for a number ofkey sources. What is immediately obvious from this plot isthat most objects have sharply decreasing CO line strengthsbeyond J > 8, in particular starforming systems such asNGC 253, or the sub-millimeter galaxy plotted in this dia-gram (the quasars appear to be more excited, but their COline SED still turns over at J ! 7, for an exceptional objectsee APM 07279, Weiss et al. 2007). This comparison im-mediately implies that emission from the CO molecule willtypically be very difficult to observe with ALMA at z > 7 asthe observable lines will simply not be excited.

2.2 Ionized carbon ([CII]) to the rescue!

An alternative tracer of the interstellar medium is one ofthe main cooling line of the ISM, the 2P3/2 " 2P1/2 fine-structure line of C+ (or [CII]). In brief, the [CII] line isexpected to be much stronger than any of the CO lines.Given its high frequency (157.74 µm, corresponding to1900.54 GHz) [CII] studies in the local universe are lim-ited to airborne or satellite missions (e.g. Stacey et al. 1991;Malhotra et al. 1997; Madden et al. 1997). These studies

Fig. 2 Comparison of various normalized (by their CO(1–0) flux den-sity) CO line SEDs at low and high redshift (figure taken from Weisset al. 2007). The CO line SEDs decline rapidly beyond J = 6–8

have demonstrated that this single line can indeed carry agood fraction of the total infrared luminosity (LFIR) of anentire galaxy. In the local universe, the ratio LCII/LFIR hasbeen found to be 2–5 # 10$4 in the case of ULIRGS (e.g.Gerin and Phillips 2000), but is more like 5–10 # 10$3 inmore typical starforming galaxies (for a discussion on pos-sible reasons for the suppressed ratio in ULIRGs see, e.g.,Luhman et al. 1998). Notably, the ratio has been found to be1% or even higher in low metallicity environments. E.g., inthe low-metallicity galaxy IC10, LCII/LFIR reaches valuesas high as 4%, with an average value of 2% (Madden et al.1997; see Israel et al. 1996 for a similar result for the LMC).This is the reason why it has long been argued (e.g., Stark1997) that observation of the [CII] line of pristine systemsat the highest redshifts will likely be the key to study mole-cular gas in the earliest starforming systems, in particularin the era of ALMA. The ALMA [CII] ‘discovery space’ isalso indicated in Fig. 1.

3 Expected [CII] line strengths

At the redshifts of the LAEs, the [CII] line is shifted to the1 mm band of ALMA (band 6, 211–275 GHz). [CII] emis-sion has recently been successfully detected using the IRAM30 m in the highest redshift quasar J1148+5251 at z = 6.42(Maiolino et al. 2005, see Fig. 3). The notable differencebetween J1148+5251 and the z > 6 LAEs is that the ratioLCII/LFIR has been found to be very low (!5 # 10$4) in


Documents

Beyond the Peak - herschel.esac.esa.intherschel.esac.esa.int/TheUniverseExploredByHerschel/presentations/10a... · Beyond the Peak Resolved Far-Infrared Spectral Mapping of Nearby